--- trunk/Estraier.pm	2006/01/06 12:40:23	49
+++ trunk/Estraier.pm	2006/01/07 00:00:15	60
@@ -4,7 +4,7 @@
 use strict;
 use warnings;
 
-our $VERSION = '0.00';
+our $VERSION = '0.01';
 
 =head1 NAME
 
@@ -205,7 +205,8 @@
 
 sub attr_names {
 	my $self = shift;
-	croak "attr_names return array, not scalar" if (! wantarray);
+	return unless ($self->{attrs});
+	#croak "attr_names return array, not scalar" if (! wantarray);
 	return sort keys %{ $self->{attrs} };
 }
 
@@ -221,8 +222,8 @@
 sub attr {
 	my $self = shift;
 	my $name = shift;
-
-	return $self->{'attrs'}->{ $name };
+	return unless (defined($name) && $self->{attrs});
+	return $self->{attrs}->{ $name };
 }
 
 
@@ -236,8 +237,8 @@
 
 sub texts {
 	my $self = shift;
-	confess "texts return array, not scalar" if (! wantarray);
-	return @{ $self->{dtexts} };
+	#confess "texts return array, not scalar" if (! wantarray);
+	return @{ $self->{dtexts} } if ($self->{dtexts});
 }
 
 
@@ -251,7 +252,7 @@
 
 sub cat_texts {
 	my $self = shift;
-	return join(' ',@{ $self->{dtexts} });
+	return join(' ',@{ $self->{dtexts} }) if ($self->{dtexts});
 }
 
 
@@ -460,7 +461,7 @@
 sub attrs {
 	my $self = shift;
 	#croak "attrs return array, not scalar" if (! wantarray);
-	return @{ $self->{attrs} };
+	return @{ $self->{attrs} } if ($self->{attrs});
 }
 
 
@@ -645,7 +646,7 @@
 
 sub doc_num {
 	my $self = shift;
-	return $#{$self->{docs}};
+	return $#{$self->{docs}} + 1;
 }
 
 
@@ -717,10 +718,10 @@
 	};
 	bless($self, $class);
 
-	if (@_) {
-		$self->{debug} = shift;
-		warn "## Node debug on\n";
-	}
+	my $args = {@_};
+
+	$self->{debug} = $args->{debug};
+	warn "## Node debug on\n" if ($self->{debug});
 
 	$self ? return $self : return undef;
 }
@@ -867,7 +868,7 @@
 	return unless ($self->{url});
 	$self->shuttle_url( $self->{url} . '/out_doc',
 		'application/x-www-form-urlencoded',
-		"uri=$uri",
+		"uri=" . uri_escape($uri),
 		undef
 	) == 200;
 }
@@ -1050,7 +1051,7 @@
 		croak "id must be numberm not '$a->{id}'" unless ($a->{id} =~ m/^\d+$/);
 		$arg = 'id=' . $a->{id};
 	} elsif ($a->{uri}) {
-		$arg = 'uri=' . $a->{uri};
+		$arg = 'uri=' . uri_escape($a->{uri});
 	} else {
 		confess "unhandled argument. Need id or uri.";
 	}
@@ -1095,7 +1096,7 @@
 
 sub name {
 	my $self = shift;
-	$self->set_info unless ($self->{name});
+	$self->_set_info unless ($self->{name});
 	return $self->{name};
 }
 
@@ -1108,7 +1109,7 @@
 
 sub label {
 	my $self = shift;
-	$self->set_info unless ($self->{label});
+	$self->_set_info unless ($self->{label});
 	return $self->{label};
 }
 
@@ -1121,7 +1122,7 @@
 
 sub doc_num {
 	my $self = shift;
-	$self->set_info if ($self->{dnum} < 0);
+	$self->_set_info if ($self->{dnum} < 0);
 	return $self->{dnum};
 }
 
@@ -1134,7 +1135,7 @@
 
 sub word_num {
 	my $self = shift;
-	$self->set_info if ($self->{wnum} < 0);
+	$self->_set_info if ($self->{wnum} < 0);
 	return $self->{wnum};
 }
 
@@ -1147,24 +1148,188 @@
 
 sub size {
 	my $self = shift;
-	$self->set_info if ($self->{size} < 0);
+	$self->_set_info if ($self->{size} < 0);
 	return $self->{size};
 }
 
 
+=head2 search
+
+Search documents which match condition
+
+  my $nres = $node->search( $cond, $depth );
+
+C<$cond> is C<Search::Estraier::Condition> object, while <$depth> specifies
+depth for meta search.
+
+Function results C<Search::Estraier::NodeResult> object.
+
+=cut
+
+sub search {
+	my $self = shift;
+	my ($cond, $depth) = @_;
+	return unless ($cond && defined($depth) && $self->{url});
+	croak "cond mush be Search::Estraier::Condition, not '$cond->isa'" unless ($cond->isa('Search::Estraier::Condition'));
+	croak "depth needs number, not '$depth'" unless ($depth =~ m/^\d+$/);
+
+	my $resbody;
+
+	my $rv = $self->shuttle_url( $self->{url} . '/search',
+		'application/x-www-form-urlencoded',
+		$self->cond_to_query( $cond ),
+		\$resbody,
+	);
+	return if ($rv != 200);
+
+	my (@docs, $hints);
+
+	my @lines = split(/\n/, $resbody);
+	return unless (@lines);
+
+	my $border = $lines[0];
+	my $isend = 0;
+	my $lnum = 1;
+
+	while ( $lnum <= $#lines ) {
+		my $line = $lines[$lnum];
+		$lnum++;
+
+		#warn "## $line\n";
+		if ($line && $line =~ m/^\Q$border\E(:END)*$/) {
+			$isend = $1;
+			last;
+		}
+
+		if ($line =~ /\t/) {
+			my ($k,$v) = split(/\t/, $line, 2);
+			$hints->{$k} = $v;
+		}
+	}
+
+	my $snum = $lnum;
+
+	while( ! $isend && $lnum <= $#lines ) {
+		my $line = $lines[$lnum];
+		#warn "# $lnum: $line\n";
+		$lnum++;
+
+		if ($line && $line =~ m/^\Q$border\E/) {
+			if ($lnum > $snum) {
+				my $rdattrs;
+				my $rdvector;
+				my $rdsnippet;
+				
+				my $rlnum = $snum;
+				while ($rlnum < $lnum - 1 ) {
+					#my $rdline = $self->_s($lines[$rlnum]);
+					my $rdline = $lines[$rlnum];
+					$rlnum++;
+					last unless ($rdline);
+					if ($rdline =~ /^%/) {
+						$rdvector = $1 if ($rdline =~ /^%VECTOR\t(.+)$/);
+					} elsif($rdline =~ /=/) {
+						$rdattrs->{$1} = $2 if ($rdline =~ /^(.+)=(.+)$/);
+					} else {
+						confess "invalid format of response";
+					}
+				}
+				while($rlnum < $lnum - 1) {
+					my $rdline = $lines[$rlnum];
+					$rlnum++;
+					$rdsnippet .= "$rdline\n";
+				}
+				#warn Dumper($rdvector, $rdattrs, $rdsnippet);
+				if (my $rduri = $rdattrs->{'@uri'}) {
+					push @docs, new Search::Estraier::ResultDocument(
+						uri => $rduri,
+						attrs => $rdattrs,
+						snippet => $rdsnippet,
+						keywords => $rdvector,
+					);
+				}
+			}
+			$snum = $lnum;
+			#warn "### $line\n";
+			$isend = 1 if ($line =~ /:END$/);
+		}
+
+	}
+
+	if (! $isend) {
+		warn "received result doesn't have :END\n$resbody";
+		return;
+	}
+
+	#warn Dumper(\@docs, $hints);
+
+	return new Search::Estraier::NodeResult( docs => \@docs, hints => $hints );
+}
+
+
+=head2 cond_to_query
+
+Return URI encoded string generated from Search::Estraier::Condition
+
+  my $args = $node->cond_to_query( $cond );
+
+=cut
+
+sub cond_to_query {
+	my $self = shift;
+
+	my $cond = shift || return;
+	croak "condition must be Search::Estraier::Condition, not '$cond->isa'" unless ($cond->isa('Search::Estraier::Condition'));
+
+	my @args;
+
+	if (my $phrase = $cond->phrase) {
+		push @args, 'phrase=' . uri_escape($phrase);
+	}
+
+	if (my @attrs = $cond->attrs) {
+		for my $i ( 0 .. $#attrs ) {
+			push @args,'attr' . ($i+1) . '=' . uri_escape( $attrs[$i] );
+		}
+	}
+
+	if (my $order = $cond->order) {
+		push @args, 'order=' . uri_escape($order);
+	}
+		
+	if (my $max = $cond->max) {
+		push @args, 'max=' . $max;
+	} else {
+		push @args, 'max=' . (1 << 30);
+	}
+
+	if (my $options = $cond->options) {
+		push @args, 'options=' . $options;
+	}
+
+	push @args, 'depth=' . $self->{depth} if ($self->{depth});
+	push @args, 'wwidth=' . $self->{wwidth};
+	push @args, 'hwidth=' . $self->{hwidth};
+	push @args, 'awidth=' . $self->{awidth};
+
+	return join('&', @args);
+}
+
 
 =head2 shuttle_url
 
 This is method which uses C<IO::Socket::INET> to communicate with Hyper Estraier node
 master.
 
-  my $rv = shuttle_url( $url, $content_type, \$req_body, \$resbody );
+  my $rv = shuttle_url( $url, $content_type, $req_body, \$resbody );
 
 C<$resheads> and C<$resbody> booleans controll if response headers and/or response
 body will be saved within object.
 
 =cut
 
+use LWP::UserAgent;
+
 sub shuttle_url {
 	my $self = shift;
 
@@ -1183,81 +1348,37 @@
 		return -1;
 	}
 
-	my ($host,$port,$query) = ($url->host, $url->port, $url->path);
-
-	if ($self->{pxhost}) {
-		($host,$port) = ($self->{pxhost}, $self->{pxport});
-		$query = "http://$host:$port/$query";
-	}
-
-	$query .= '?' . $url->query if ($url->query && ! $reqbody);
-
-	my $headers;
+	my $ua = LWP::UserAgent->new;
+	$ua->agent( "Search-Estraier/$Search::Estraier::VERSION" );
 
+	my $req;
 	if ($reqbody) {
-		$headers .= "POST $query HTTP/1.0\r\n";
+		$req = HTTP::Request->new(POST => $url);
 	} else {
-		$headers .= "GET $query HTTP/1.0\r\n";
+		$req = HTTP::Request->new(GET => $url);
 	}
 
-	$headers .= "Host: " . $url->host . ":" . $url->port . "\r\n";
-	$headers .= "Connection: close\r\n";
-	$headers .= "User-Agent: Search-Estraier/$Search::Estraier::VERSION\r\n";
-	$headers .= "Content-Type: $content_type\r\n";
-	$headers .= "Authorization: Basic $self->{auth}\r\n";
-	my $len = 0;
-	{
-		use bytes;
-		$len = length($reqbody) if ($reqbody);
-	}
-	$headers .= "Content-Length: $len\r\n";
-	$headers .= "\r\n";
-
-	my $sock = IO::Socket::INET->new(
-		PeerAddr	=> $host,
-		PeerPort	=> $port,
-		Proto		=> 'tcp',
-		Timeout		=> $self->{timeout} || 90,
-	);
+	$req->headers->header( 'Host' => $url->host . ":" . $url->port );
+	$req->headers->header( 'Connection', 'close' );
+	$req->headers->header( 'Authorization', 'Basic ' . $self->{auth} );
+	$req->content_type( $content_type );
 
-	if (! $sock) {
-		carp "can't open socket to $host:$port";
-		return -1;
-	}
-
-	warn $headers if ($self->{debug});
-
-	print $sock $headers or
-		carp "can't send headers to network:\n$headers\n" and return -1;
+	warn $req->headers->as_string,"\n" if ($self->{debug});
 
 	if ($reqbody) {
 		warn "$reqbody\n" if ($self->{debug});
-		print $sock $reqbody or
-			carp "can't send request body to network:\n$$reqbody\n" and return -1;
+		$req->content( $reqbody );
 	}
 
-	my $line = <$sock>;
-	chomp($line);
-	my ($schema, $res_status, undef) = split(/  */, $line, 3);
-	return if ($schema !~ /^HTTP/ || ! $res_status);
-
-	$self->{status} = $res_status;
-	warn "## response status: $res_status\n" if ($self->{debug});
-
-	# skip rest of headers
-	$line = <$sock>;
-	while ($line) {
-		$line = <$sock>;
-		$line =~ s/[\r\n]+$//;
-		warn "## ", $line || 'NULL', " ##\n" if ($self->{debug});
-	};
+	my $res = $ua->request($req) || croak "can't make request to $url: $!";
+
+	warn "## response status: ",$res->status_line,"\n" if ($self->{debug});
 
-	# read body
-	$len = 0;
-	do {
-		$len = read($sock, my $buf, 8192);
-		$$resbody .= $buf if ($resbody);
-	} while ($len);
+	return -1 if (! $res->is_success);
+
+	($self->{status}, $self->{status_message}) = split(/\s+/, $res->status_line, 2);
+
+	$$resbody .= $res->content;
 
 	warn "## response body:\n$$resbody\n" if ($resbody && $self->{debug});
 
@@ -1265,15 +1386,117 @@
 }
 
 
-=head2 set_info
+=head2 set_snippet_width
+
+Set width of snippets in results
+
+  $node->set_snippet_width( $wwidth, $hwidth, $awidth );
+
+C<$wwidth> specifies whole width of snippet. It's C<480> by default. If it's C<0> snippet
+is not sent with results. If it is negative, whole document text is sent instead of snippet.
+
+C<$hwidth> specified width of strings from beginning of string. Default
+value is C<96>. Negative or zero value keep previous value.
+
+C<$awidth> specifies width of strings around each highlighted word. It's C<96> by default.
+If negative of zero value is provided previous value is kept unchanged.
+
+=cut
+
+sub set_snippet_width {
+	my $self = shift;
+
+	my ($wwidth, $hwidth, $awidth) = @_;
+	$self->{wwidth} = $wwidth;
+	$self->{hwidth} = $hwidth if ($hwidth >= 0);
+	$self->{awidth} = $awidth if ($awidth >= 0);
+}
+
+
+=head2 set_user
+
+Manage users of node
+
+  $node->set_user( 'name', $mode );
+
+C<$mode> can be one of:
+
+=over 4
+
+=item 0
+
+delete account
+
+=item 1
+
+set administrative right for user
+
+=item 2
+
+set user account as guest
+
+=back
+
+Return true on success, otherwise false.
+
+=cut
+
+sub set_user {
+	my $self = shift;
+	my ($name, $mode) = @_;
+
+	return unless ($self->{url});
+	croak "mode must be number, not '$mode'" unless ($mode =~ m/^\d+$/);
+
+	$self->shuttle_url( $self->{url} . '/_set_user',
+		'text/plain',
+		'name=' . uri_escape($name) . '&mode=' . $mode,
+		undef
+	) == 200;
+}
+
+
+=head2 set_link
+
+Manage node links
+
+  $node->set_link('http://localhost:1978/node/another', 'another node label', $credit);
+
+If C<$credit> is negative, link is removed.
+
+=cut
+
+sub set_link {
+	my $self = shift;
+	my ($url, $label, $credit) = @_;
+
+	return unless ($self->{url});
+	croak "mode credit be number, not '$credit'" unless ($credit =~ m/^\d+$/);
+
+	my $reqbody = 'url=' . uri_escape($url) . '&label=' . uri_escape($label);
+	$reqbody .= '&credit=' . $credit if ($credit > 0);
+
+	$self->shuttle_url( $self->{url} . '/_set_link',
+		'text/plain',
+		$reqbody,
+		undef
+	) == 200;
+}
+
+
+=head1 PRIVATE METHODS
+
+You could call those directly, but you don't have to. I hope.
+
+=head2 _set_info
 
 Set information for node
 
-  $node->set_info;
+  $node->_set_info;
 
 =cut
 
-sub set_info {
+sub _set_info {
 	my $self = shift;
 
 	$self->{status} = -1;
@@ -1288,7 +1511,8 @@
 
 	return if ($rv != 200 || !$resbody);
 
-	chomp($resbody);
+	# it seems that response can have multiple line endings
+	$resbody =~ s/[\r\n]+$//;
 
 	( $self->{name}, $self->{label}, $self->{dnum}, $self->{wnum}, $self->{size} ) =
 		split(/\t/, $resbody, 5);